@jon I do appreciate this perspective as well. Perhaps it’s about finding the right words. I obviously put these thoughts out there to gather how others reason. I’ve been reacting a lot lately to expressions of “good” use disregarding any second or third order thinking.
Not sure I would agree it’s absolutist though, for me it’s about choosing to weigh in impact beyond immediate personal benefits and reasoning around them.
The question of “useful” always for me ends up at needing to understand useful for who, at who’s expense? And from there: is it worth it? Perhaps, for many it, is. A great many people of course have no awareness at all of the harm, in my experience.
I’d also emphasize that I am adressing the modern flavor of generative AI, rather than all the many flavors of machine automation. For translation, transcription and similar applications – within a dominant language – I’d argue that large language models are incredibly useful. For many other applications it’s both deceptive and harmful, even ones where it’s argued they provide value. And so, I have to reason around if the positive applications are worth the many and varied costs, and hence can be considered good applications for the goals they help achieve (goals which likely can be arrived at in many other ways as well).
Clearly I’m concerned when the primary beneficiaries are those who are already well-off, and the sufferers are those who are subjected to exploitation time and time again. Much of the generative AI world amplifies this arrangement. Perhaps it’s unfair of me to judge someone’s experience of useful and good under this light, but I’m having a really hard time accepting those conclusions when they risk exacerbating harm…
I’m going to give this some more thought and work it into some writing on my blog… thanks for offering objection, and sorry about a wall of unfinished pondering - much of how I reasoned was perhaps obvious.